Can civic engagement programs discourage police and party activists from engaging in electoral violence?
EGAP researchers: Sabrina Karim
Other researchers: Lindsey Pruett, Alex Dyzenhaus, and Dao Freeman
Key takeaway: Direct experience with well-managed post-conflict elections is associated with improved attitudes toward democracy and reduced support for electoral violence among police and party activists, even in settings where targeted, locally-led civic engagement programs may struggle to alter these attitudes.
Geographical Region: Africa
Type of study: Field experiment
Preparer: Mark Williamson
Executive Summary
Civic engagement programs have been proposed as a potential intervention to reduce electoral violence in post-conflict settings. Yet these interventions are rarely tested with the actors who are most likely to be involved in this type of violence: police and party activists. This study reports the results of a field experiment involving a civic engagement program with participants from these groups during Liberia’s 2017 election. While the program did not have significant effects on the attitudes of the police or activists, the authors do find that both groups’ views on electoral violence, democracy and each other improved after the country’s first ever election without third-party oversight. These findings highlight how direct experience with a well-managed, independent and nonviolent election can reinforce democratic values during crucial post-conflict moments.
Policy Challenge
In many post-conflict democracies, electoral violence is a persistent threat to stability. Police and youth-wing political party activists often perpetrate low-level violence alongside their participation in basic electoral processes, like campaigning and security provision. While these actors’ choice to engage in violence is informed by directives from political elites, they often also maintain a significant degree of autonomy. Police and activists have to make day-to-day decisions, for example, about how to respond to protests, threats made at polling stations, or perceived irregularities in the electoral process. In settings where there is significant uncertainty—about how others will behave, about the acceptability of violence, or about nonviolent means for resolving disputes—these actors may resort to violence.
Civic engagement programs have been proposed as one way to address this uncertainty and, by extension, reduce the potential for electoral violence. These programs provide information on violence and its prevention, legal procedures for addressing conflict and democratic processes. Through interaction between potentially opposing groups in a neutral space, they can also reduce misperceptions about others’ intentions. However, most studies into the effectiveness of civic engagement programming focus on ordinary civilians, rather than those who are most likely to engage in electoral violence: police and party activists. This study helps address this evidence gap through a field experiment in Liberia specifically focused on these actors.
Context
From 1989 to 2003, Liberia experienced two civil wars. Its first two post-conflict elections, in 2005 and 2011, saw low-intensity violence perpetrated by police and youth party activists. In 2017, Liberia held a crucial election: the first vote in which the electoral process was not overseen or managed by external actors. After a significant scale back of the UN peacekeeping presence in 2016, this represented a first real test of the country’s democratic institutions.
At this time, a former officer with the Liberian National Police (LNP) developed a civic engagement program to improve relationships between police and youth party activists through small informational workshops, mentorship meetings and dialogue. This program, implemented through a local non-governmental organization (NGO), aimed to improve understanding between the two groups and teach conflict resolution strategies.
Research Design
To test the effectiveness of this civic engagement program, a sample of 120 active police officers and 300 youth party members from all political parties was recruited. After completing an initial baseline survey, 70 and 200 participants from each group were randomly selected to participate in the program. The treatment group engaged in a series of dialogue and mentorship meetings in the months preceding the 2017 election, as well as one follow-up meeting afterwards.
Several attitudinal outcomes were measured in an endline survey after the program ended. These outcomes were aggregated into four indices corresponding to the following topics: (1) attitudes towards violence; (2) knowledge about institutions and regulations; (3) views of democracy; and (4) intergroup police-youth attitudes. The questions included realistic, locally-relevant scenarios based on the program’s discussions and educational materials.
Because these same outcomes were also measured in the baseline survey, the authors are able to use a difference-in-differences design to compare how treated participants’ attitudes changed before and after the program, relative to changes in the control group over that same period.
Results
The authors find that participation in the civic engagement program did not have meaningful effects on youth or police attitudes across any of the outcome indices. Relative to those that didn’t complete the program, treated youth and police did not improve their attitudes toward violence, support for democracy, knowledge of legal dispute resolution mechanisms, or views of the police-youth out-group. These null results do not seem to be driven by model specification, non-compliance, ceiling effects or social desirability bias.
But this is not the full story. Between the baseline and endline surveys, both treated and control participants had direct experiences with Liberia’s first post-conflict election conducted without external oversight. This election revealed crucial new information about institutional capacity and democratic norms: the LNP handled election security without significant violence and opposing political parties peacefully transferred power for the first time since the civil wars.
How did exposure to these events shape participants’ attitudes? The authors answer this question by comparing the same outcomes before and after the election among both treated and control respondents. They found that over this period, all respondents came to view violence as a less appropriate electoral strategy, held better opinions of each other and became more supportive of democracy. Answers to supplementary questions about the perceived integrity of the election suggest that these improvements were largely driven by a sense that this election was more credible than previous contests.
Lessons
This study highlights two key takeaways for policymakers. First, the results emphasize how successfully administered elections—particularly at crucial post-conflict junctures—can strengthen democratic norms, even among actors that are more prone to electoral violence. Bolstering young democracies’ institutional capacity during these critical moments allows them to capitalize on citizens’ high expectations and thus reinforce democratization. These results are likely to generalize to other settings, like Liberia, where uncertainty is widespread and violence is not entirely directed by elites. Second, this study demonstrates a viable model for collaborating with local NGOs without undermining their vision and contextual knowledge. Policymakers should prioritize funding for similar locally-led research initiatives, which often lack resources to scale up their interventions to a level that smaller experimental effects can be detected with confidence.